This application is for support of genetic, biochemical, and cell biological analyses of glycosyl-phosphatidylinositols (GPIs). These glycolipids, which serve as membrane anchors for cell surface proteins, but can exist as free lipids as well, are essential for the growth of unicellular eukaryotes and mammals alike. GPI-containing proteins are key determinants of the virulence of the pathogenic fungi that infect immunocompromised patients, and GPI derivatives are also the major surface components of tropical parasites. In mammals, GPI-anchored proteins serve as toxin receptors and regulators in the immune system. Efforts will be aimed at understanding the biochemistry and cell biology of GPI synthesis and at identifying steps in GPI formation in which human and microbial cells differ. Yeast mutants defective in different stages of GPI assembly will be used as tools. Structural analyses of the GPIs that accumulate in GPI synthesis mutants reveal that they are not formed in a single linear pathway, rather, that the synthetic pathway is branched. Further, it is hypothesized that the different branches are defined by the positions at which phosphoethanolamine side chains are added to the GPIs. These novel aspects of GPI formation will be tested and explored further by determining the structures of the GPIs that accumulate in single and double mutants defective in the enzymes that transfer and phosphoethanolamine. One explanation for pathway branching is that GPI assembly occurs in separate membrane compartments that have over lapping populations of GPI synthetic enzymes but differ in their phosphoethanolamine transferases. This will be tested in subcellular fractionation and immunolocalization experiments with mammalian cells: those GPIs synthetic proteins specific to separate branches of the pathway may differ in their cellular localization. It is postulated that enzymes involved in mannose and phosphoethanolamine addition to GPIs act in membrane-bound complexes, either with other GPI synthetic proteins, or as homodimers or oligomers. Individual proteins will be tested for their ability to co-precipitate further proteins, whose identify will be established by mass spectroscopy. Co-purifying proteins may include proteins of unknown function, whose role in GPI assembly will be tested by creating mutant yeast strains deficient in them. To identify further genes involved in GPI assembly, or cellular processes dependent on GPI synthesis, genetic screens will be carried our for mutants synthetically lethal with known gpi mutants. Detailed searches of genome sequence databases have also identified genes encoding candidate GPI synthetic enzymes, whose role in GPI assembly will be explored by creating yeast mutants in them.